Vessel Risk Advisory Using ML Correlation to Cut False Alarms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Operators of marine vessels face challenges in managing the overwhelming number of variables and parameters of vessel systems, leading to reactionary responses to potential failures due to limited displayable indicators and environmental impacts, necessitating a more proactive monitoring system.
Innovation Solution
A conditional online-based risk advisory system (COBRAS) utilizing machine learning to monitor vessel systems, reduce false notifications, and provide proactive alerts and optimizations by learning from user interactions to correlate relevant variables.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional sensor-based monitoring systems are used to track vessel system parameters, then system safety is improved through detection of abnormal conditions, but the number of false notifications and alarms increases due to the overwhelming number of variables and environmental impacts
Solution Approach 1:
The patent introduces an intermediary processing layer between sensors and operators that uses machine learning models to analyze sensor data, correlate multiple variables, and filter out false alarms caused by environmental factors. This intermediary intelligence layer processes the overwhelming number of variables and presents only meaningful alerts to operators, reducing false notifications while maintaining system safety monitoring.
Solution Approach 2:
The system implements feedback mechanisms where operator responses to notifications are used to continuously improve the machine learning models. The system learns from operator corrections and adjustments, refining its ability to distinguish true anomalies from false alarms, thereby reducing false notifications over time while maintaining reliable safety monitoring.
2Measurement precision
If operators manually monitor all vessel system parameters, then detection precision is improved through direct observation, but operator workload and response time worsen due to the overwhelming number of variables
Solution Approach 1:
The system enables self-service monitoring where the machine learning model automatically performs data correlation, anomaly detection, and alert generation without requiring manual operator analysis of each parameter. The system serves itself by autonomously processing the overwhelming number of variables and presenting synthesized information to operators, maintaining detection precision while dramatically reducing operator workload.
Solution Approach 2:
The system performs preliminary analysis and filtering of vessel parameters before presenting information to operators. Machine learning models pre-process sensor data, correlate variables, and identify potential issues in advance, so operators receive pre-analyzed, prioritized alerts rather than raw data requiring manual interpretation, thereby reducing workload while maintaining detection precision.
3Loss of time
If traditional alarm systems notify operators of all parameter threshold crossings, then response time is improved through immediate notification, but the number of false alarms increases due to environmental factors and variable correlations
Solution Approach 1:
An intermediary intelligence layer is introduced between parameter threshold crossing and operator notification. This layer uses machine learning to analyze the context of threshold crossings, correlate with environmental factors and other variables, and determine whether the threshold crossing represents a true alarm condition or a false alarm, thereby reducing false alarms while maintaining rapid response time for genuine issues.
Solution Approach 2:
The system performs preliminary anti-action by preemptively filtering out false alarms before they reach operators. Machine learning models predict and block false alarm notifications by identifying patterns associated with environmental factors and non-critical variations, allowing the system to prevent harmful false alarm notifications while maintaining rapid response to genuine threats.
Data Source
AI summary
An advisory system of a vessel that monitors variables of a vessel system inclusive of systems and subsystems that are used to operate the vessel. The advisory system may use machine-learning to learn from an operator (i) whether or not two variables are related to one another, and (ii) likelihood that a variable will reach a threshold, and, optionally, time until reaching the threshold. The system may receive operator feedback (i) to indicate whether the two variables are related to one another, and (ii) whether a behavior of the variable is normal or not normal. Thereafter, if a determination that the same two variables are related to one another and behaving in a similar manner, provide notification to the operator of the behavior. In response to determining that the variable is behaving (e.g., trending) in a similar manner that is not normal, providing a notification to the operator.